Computer history: Graphical systems (1978 – 1983)
Graphical Systems (1978 – 1983)
PDP 11/60 and Evans & Sutherland PS/2
In mid-1979, the CDC Digigraphics graphics computer system was decommissioned. Instead, a DEC PDP 11/60 – an Evans & Sutherland Picture System/2 (PS/2) combination was installed. The PS/2 system cost about US$ 110,000, and the PDP 11/60 system cost about the same. The PDP 11/60 system contained 192 KB of memory, a CPU with a hardware floating point unit and four interchangeable RL01 disks with a storage capacity of 5 MB each.
The installation was planned for April 1979. A major delay occurred as heavy items were placed on top of both systems during the transport by plane. The DEC 11/60 system frame was distorted. Moreover, a large cathode ray tube had tumbled from the aircraft during unloading. The tube had shot more than five centimetres back into the housing frame. It was still working! Eventually, a replacement PDP 11/60 system was delivered at the end of July 1979 and the other damages were repaired.
The Laboratory tried to get an import tax exemption for the PS/2 graphics system as it was considered an “object of scientific nature mainly suitable for executing scientific activities”. This request was rejected by Dutch Customs. In 1980 the time had come that minicomputers and graphics systems became ordinary off-the-shelf devices.
At the end of 1980, the Picture System/2 was extended to a Multi-Picture System. The main user of the system was the research group Telecommunications, which at the time was conducting research into the optimal design of the Zodiac communication system grid of the Dutch First Army Corps (1LK). ZODIAC is short in Dutch for ZOne, DIgital, Automated and Cryptographically protected. Among other things, TNO investigated how communication nodes in the field could be delinked, moved and reconnected without reachability impact on the remainder of the Zodiac grid. A description of Zodiac stated: “Mobility and a higher movement frequency were essential for survivability and responding to changes in tactical situations.”
The CDC Cyber 18-17: plotter and D-Manager
The CDC Cyber 18-17 (shortly System 17), which had controlled the Digigraphic earlier, was reused as a minicomputer to control the Calcomp 936-plotter and to host the DManager. The D(ata) Manager was an own Laboratory development: a bus structure to which up to 63 processors (POPs) could be linked. In addition to “Janus”, the CDC CYBER PP-program 1IR which controlled both the card reader, printers and the Calcomp plotter, a special PP overlay was developed that made a ‘hand-shake’ via a channel coupling between the CDC 6400 and the Cyber 18-17. Subsequently, the plot data files were transferred from the output queue on the 6400 to the Cyber 18-17, which in turn controlled the plotter. The channel speed was about 3 to 4 Mbyte/s. After plotting, the plot files were stored for some time on the CDC 6400 to be able to output the plot again if, for example, an ink pen had dried out.
In the seventies, Tektronix manufactured the 4010 and the larger 4014 graphics terminals. “Drawings” and text could be displayed on phosphor tubes. The Laboratory had some of those smaller terminals and installed one large TEK4014 terminal with a hard-copy print unit in the terminal room at the end of the 1970s. These terminals could be controlled using the routines from the Terminal Control System (TCS) library (PLOT 10). The visual screen size was 1023 x 780 pixels.
As the Calcomp plotter had been in use for more than ten years and showed early signs of wear and tear, we replaced it with a Calcomp 1051 plotter in 1982. Because the plotter had a high percentage of utilisation, there was a great demand for a plot-preview possibility. After studying the Calcomp plot library, it took us only a few weeks to program a library that replaced the Calcomp plot library calls with TCS/PLOT 10 calls (at the bottom). As a result, users could preview their plots and remove program and dataset errors at a Tektronix 4010 or 4014 terminal. The PREVIEW program retrieved the preview plot from a system queue, and showed it, after which the user could “zoom” in. When he/she was really satisfied, the plot could be transferred to the plot queue with a simple click.
The Comtal “Vision One/20” system
The COMTAL Vision One/20 – according to the supplier “a system that explored the edges of the advanced technical possibilities” was delivered in October 1980. The purchase costs amounted to US$ 144,000. Upon acceptance, the system did not meet all the previously agreed requirements. Because of project obligations, the system was conditionally accepted. A certain amount of the purchase price was withheld until the problems were resolved. The supplied Comtal Vision One/20 was based on a PDP LSI 11/02 system. To solve the acceptance test problems, we had to switch to Release K, a release which did not run on a PDP 11/02. Some months later, our supplier broke its commercial links with the manufacturer of the system.
In 1987- seven years after the installation – some attorneys still tried to settle the purchase and payment conflict. We decommissioned the Comtal system on 1 September 1988.
In 1983, we purchased a VAX 11/750 for about 350,000 guilders to:
- control the Comtal and allow the storage of Comtal files to be able to use the Comtal more effectively;
- develop software for measuring systems based on microPower/Pascal;
- cross-compile own developed code for Motorola 68000 microprocessors, especially for the Defence Zodiac communication systems.
When the VAX 11/750 was purchased, the Computer Committee of the TNO Defence Research Group (HDO), the forerunner of TNO Defence Research, looked explicitly at the feasibility of the system in the infrastructure of the ‘integrated’ laboratory as well as that of the HDO.